source: src/Actions/FragmentationAction/FragmentationAutomationAction.cpp@ adccae

/*
 * Project: MoleCuilder
 * Description: creates and alters molecular systems
 * Copyright (C)  2010-2012 University of Bonn. All rights reserved.
 * 
 *
 *   This file is part of MoleCuilder.
 *
 *    MoleCuilder is free software: you can redistribute it and/or modify
 *    it under the terms of the GNU General Public License as published by
 *    the Free Software Foundation, either version 2 of the License, or
 *    (at your option) any later version.
 *
 *    MoleCuilder is distributed in the hope that it will be useful,
 *    but WITHOUT ANY WARRANTY; without even the implied warranty of
 *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *    GNU General Public License for more details.
 *
 *    You should have received a copy of the GNU General Public License
 *    along with MoleCuilder.  If not, see <http://www.gnu.org/licenses/>.
 */

/*
 * FragmentationAutomationAction.cpp
 *
 *  Created on: May 18, 2012
 *      Author: heber
 */

// include config.h
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif

#include <boost/archive/text_iarchive.hpp>
// boost asio needs specific operator new
#include <boost/asio.hpp>

#include "CodePatterns/MemDebug.hpp"

#include <boost/assign.hpp>

#include "CodePatterns/Assert.hpp"
#include "CodePatterns/Info.hpp"
#include "CodePatterns/Log.hpp"
#include "JobMarket/Controller/FragmentController.hpp"
#include "JobMarket/Jobs/FragmentJob.hpp"

#include "Atom/atom.hpp"
#include "Fragmentation/EnergyMatrix.hpp"
#include "Fragmentation/ForceMatrix.hpp"
#include "Fragmentation/Fragmentation.hpp"
#include "Fragmentation/Histogram/Histogram.hpp"
#include "Fragmentation/HydrogenSaturation_enum.hpp"
#include "Fragmentation/KeySet.hpp"
#include "Fragmentation/KeySetsContainer.hpp"
#include "Fragmentation/Summation/Summator.hpp"
#include "Graph/DepthFirstSearchAnalysis.hpp"
#include "Jobs/MPQCJob.hpp"
#include "Jobs/MPQCData.hpp"
#include "Jobs/MPQCData_printName.hpp"
#include "molecule.hpp"
#include "World.hpp"

#include <iostream>
#include <string>
#include <vector>

#include <boost/mpl/for_each.hpp>

#include "Actions/FragmentationAction/FragmentationAutomationAction.hpp"

using namespace MoleCuilder;

using namespace boost::assign;

// and construct the stuff
#include "FragmentationAutomationAction.def"
#include "Action_impl_pre.hpp"
/** =========== define the function ====================== */

class controller_AddOn;

// needs to be defined for using the FragmentController
controller_AddOn *getAddOn()
{
  return NULL;
}

/** Creates a MPQCCommandJob with argument \a filename.
 *
 * @param jobs created job is added to this vector
 * @param command mpqc command to execute
 * @param filename filename being argument to job
 * @param nextid id for this job
 */
void parsejob(
    std::vector<FragmentJob::ptr> &jobs,
    const std::string &command,
    const std::string &filename,
    const JobId_t nextid)
{
  std::ifstream file;
  file.open(filename.c_str());
  ASSERT( file.good(), "parsejob() - file "+filename+" does not exist.");
  std::string output((std::istreambuf_iterator<char>(file)),
      std::istreambuf_iterator<char>());
  FragmentJob::ptr testJob( new MPQCJob(nextid, output) );
  jobs.push_back(testJob);
  file.close();
  LOG(1, "INFO: Added MPQCCommandJob from file "+filename+".");
}

/** Helper function to get number of atoms somehow.
 *
 * Here, we just parse the number of lines in the adjacency file as
 * it should correspond to the number of atoms, except when some atoms
 * are not bonded, but then fragmentation makes no sense.
 *
 * @param path path to the adjacency file
 */
size_t getNoAtomsFromAdjacencyFile(const std::string &path)
{
  size_t NoAtoms = 0;

  // parse in special file to get atom count (from line count)
  std::string filename(path);
  filename += FRAGMENTPREFIX;
  filename += ADJACENCYFILE;
  std::ifstream adjacency(filename.c_str());
  if (adjacency.fail()) {
    LOG(0, endl << "getNoAtomsFromAdjacencyFile() - Unable to open " << filename << ", is the directory correct?");
    return false;
  }
  std::string buffer;
  while (getline(adjacency, buffer))
    NoAtoms++;
  LOG(1, "INFO: There are " << NoAtoms << " atoms.");

  return NoAtoms;
}


template <typename MapType>
struct printSum {
  printSum(
      SubsetMap::ptr &_subsetmap,
      const std::vector<MapType> &_data,
      const std::vector<JobId_t> &_jobids,
      const IndexSetContainer::Container_t &_container,
      std::map< JobId_t, size_t > &_MatrixNrLookup) : /* cannot make this const due to operator[] */
    subsetmap(_subsetmap),
    data(_data),
    jobids(_jobids),
    container(_container),
    MatrixNrLookup(_MatrixNrLookup)
  {}
  template <typename MapKey>
  void operator()(MapKey &) {
    Summator<MapType, MapKey> sum_value(
        subsetmap, data, jobids, container, MatrixNrLookup
        );
    double value = sum_value();
    LOG(0, "STATUS: Resulting " << MPQCDataNames::printName<MapKey>() << " is " << value << ".");
  }

private:
  //!> Hierarchy of IndexSet's
  SubsetMap::ptr &subsetmap;
  //!> vector of data converted from MPQCData
  const std::vector<MapType> &data;
  //!> vector of jobids
  const std::vector<JobId_t> &jobids;
  //!> container with all IndexSet's
  const IndexSetContainer::Container_t &container;
  //!> lookup map from job ids to ordering in above vectors
  std::map< JobId_t, size_t > MatrixNrLookup;
};

/** Print MPQCData from received results.
 *
 * @param results received results to extract MPQCData from
 * @param KeySetFilename filename with keysets to associate forces correctly
 * @param NoAtoms total number of atoms
 */
bool printReceivedMPQCResults(
    const std::vector<FragmentResult::ptr> &results,
    const std::string &KeySetFilename,
    size_t NoAtoms)
{
  EnergyMatrix Energy;
  EnergyMatrix EnergyFragments;
  ForceMatrix Force;
  ForceMatrix ForceFragments;

  // align fragments
  std::map< JobId_t, size_t > MatrixNrLookup;
  size_t FragmentCounter = 0;
  {
    // bring ids in order ...
    typedef std::map< JobId_t, FragmentResult::ptr> IdResultMap_t;
    IdResultMap_t IdResultMap;
    for (std::vector<FragmentResult::ptr>::const_iterator iter = results.begin();
        iter != results.end(); ++iter) {
  #ifndef NDEBUG
      std::pair< IdResultMap_t::iterator, bool> inserter =
  #endif
      IdResultMap.insert( make_pair((*iter)->getId(), *iter) );
      ASSERT( inserter.second,
          "printReceivedMPQCResults() - two results have same id "
          +toString((*iter)->getId())+".");
    }
    // ... and fill lookup
    for(IdResultMap_t::const_iterator iter = IdResultMap.begin();
        iter != IdResultMap.end(); ++iter)
      MatrixNrLookup.insert( make_pair(iter->first, FragmentCounter++) );
  }
  LOG(1, "INFO: There are " << FragmentCounter << " fragments.");

  // extract results
  std::vector<MPQCData> fragmentData(results.size());
  MPQCData combinedData;

  LOG(2, "DEBUG: Parsing now through " << results.size() << " results.");
  for (std::vector<FragmentResult::ptr>::const_iterator iter = results.begin();
      iter != results.end(); ++iter) {
    LOG(1, "RESULT: job #"+toString((*iter)->getId())+": "+toString((*iter)->result));
    MPQCData extractedData;
    std::stringstream inputstream((*iter)->result);
    LOG(2, "DEBUG: First 50 characters FragmentResult's string: "+(*iter)->result.substr(0, 50));
    boost::archive::text_iarchive ia(inputstream);
    ia >> extractedData;
    LOG(1, "INFO: extracted data is " << extractedData << ".");

    // place results into EnergyMatrix ...
    {
      MatrixContainer::MatrixArray matrix;
      matrix.resize(1);
      matrix[0].resize(1, extractedData.energies.total);
      if (!Energy.AddMatrix(
          std::string("MPQCJob ")+toString((*iter)->getId()),
          matrix,
          MatrixNrLookup[(*iter)->getId()])) {
        ELOG(1, "Adding energy matrix failed.");
        return false;
      }
    }
    // ... and ForceMatrix (with two empty columns in front)
    {
      MatrixContainer::MatrixArray matrix;
      const size_t rows = extractedData.forces.size();
      matrix.resize(rows);
      for (size_t i=0;i<rows;++i) {
        const size_t columns = 2+extractedData.forces[i].size();
        matrix[i].resize(columns, 0.);
  //      for (size_t j=0;j<2;++j)
  //        matrix[i][j] = 0.;
        for (size_t j=2;j<columns;++j)
          matrix[i][j] = extractedData.forces[i][j-2];
      }
      if (!Force.AddMatrix(
          std::string("MPQCJob ")+toString((*iter)->getId()),
          matrix,
          MatrixNrLookup[(*iter)->getId()])) {
        ELOG(1, "Adding force matrix failed.");
        return false;
      }
    }
  }
  // add one more matrix (not required for energy)
  MatrixContainer::MatrixArray matrix;
  matrix.resize(1);
  matrix[0].resize(1, 0.);
  if (!Energy.AddMatrix(std::string("MPQCJob total"), matrix, FragmentCounter))
    return false;
  // but for energy because we need to know total number of atoms
  matrix.resize(NoAtoms);
  for (size_t i = 0; i< NoAtoms; ++i)
    matrix[i].resize(2+NDIM, 0.);
  if (!Force.AddMatrix(std::string("MPQCJob total"), matrix, FragmentCounter))
    return false;

  // initialise indices
  KeySetsContainer KeySet;
  if (!Energy.InitialiseIndices()) return false;

  if (!Force.ParseIndices(KeySetFilename.c_str())) return false;

  if (!KeySet.ParseKeySets(KeySetFilename.c_str(), Force.RowCounter, Force.MatrixCounter)) return false;

  /// prepare for OrthogonalSummation

  // gather all present indices in AllIndices
  IndexSet::ptr AllIndices(new IndexSet);
  for (KeySetsContainer::ArrayOfIntVectors::const_iterator iter = KeySet.KeySets.begin();
      iter != KeySet.KeySets.end(); ++iter)
    for(KeySetsContainer::IntVector::const_iterator keyiter = (*iter).begin();
        keyiter != (*iter).end(); ++keyiter) {
      if (*keyiter != -1)
        (*AllIndices) += *keyiter;
    }
  LOG(1, "INFO: AllIndices is " << AllIndices << ".");
  // create container with all keysets
  IndexSetContainer::ptr container(new IndexSetContainer(AllIndices));
  for (KeySetsContainer::ArrayOfIntVectors::const_iterator iter = KeySet.KeySets.begin();
      iter != KeySet.KeySets.end(); ++iter) {
    IndexSet tempset;
    for(KeySetsContainer::IntVector::const_iterator keyiter = (*iter).begin();
        keyiter != (*iter).end(); ++keyiter)
      if (*keyiter != -1)
        tempset += *keyiter;
    container->insert(tempset);
  }
  // create the map of all keysets
  SubsetMap::ptr subsetmap(new SubsetMap(*container));

  {
    /// convert all MPQCData to MPQCDataMap_t

    // energy_t
    std::vector<MPQCDataEnergyMap_t> MPQCData_Energy_fused;
    MPQCData_Energy_fused.reserve(fragmentData.size());
    for(std::vector<MPQCData>::const_iterator dataiter = fragmentData.begin();
        dataiter != fragmentData.end(); ++dataiter) {
      const MPQCData &extractedData = *dataiter;
      LOG(2, "DEBUG: Current extracted Data is " << extractedData << ".");
      MPQCDataEnergyMap_t instance;
      boost::fusion::at_key<MPQCDataFused::energy_total>(instance) = extractedData.energies.total;
      boost::fusion::at_key<MPQCDataFused::energy_nuclear_repulsion>(instance) = extractedData.energies.nuclear_repulsion;
      boost::fusion::at_key<MPQCDataFused::energy_electron_coulomb>(instance) = extractedData.energies.electron_coulomb;
      boost::fusion::at_key<MPQCDataFused::energy_electron_exchange>(instance) = extractedData.energies.electron_exchange;
      boost::fusion::at_key<MPQCDataFused::energy_correlation>(instance) = extractedData.energies.correlation;
      boost::fusion::at_key<MPQCDataFused::energy_overlap>(instance) = extractedData.energies.overlap;
      boost::fusion::at_key<MPQCDataFused::energy_kinetic>(instance) = extractedData.energies.kinetic;
      boost::fusion::at_key<MPQCDataFused::energy_hcore>(instance) = extractedData.energies.hcore;
  //    boost::fusion::at_key<MPQCDataFused::energy_eigenvalues>(instance) = extractedData.energies.eigenvalues;
      MPQCData_Energy_fused.push_back(instance);
    }

    // forces
    std::vector<MPQCDataForceMap_t> MPQCData_Force_fused;
    MPQCData_Force_fused.reserve(fragmentData.size());
    for(std::vector<MPQCData>::const_iterator dataiter = fragmentData.begin();
        dataiter != fragmentData.end(); ++dataiter) {
      const MPQCData &extractedData = *dataiter;
      LOG(2, "DEBUG: Current extracted Data is " << extractedData << ".");
      MPQCDataForceMap_t instance;
      boost::fusion::at_key<MPQCDataFused::forces>(instance) = extractedData.forces;
      MPQCData_Force_fused.push_back(instance);
    }

    // times
    std::vector<MPQCDataTimeMap_t> MPQCData_Time_fused;
    MPQCData_Time_fused.reserve(fragmentData.size());
    for(std::vector<MPQCData>::const_iterator dataiter = fragmentData.begin();
        dataiter != fragmentData.end(); ++dataiter) {
      const MPQCData &extractedData = *dataiter;
      LOG(2, "DEBUG: Current extracted Data is " << extractedData << ".");
      MPQCDataTimeMap_t instance;
      boost::fusion::at_key<MPQCDataFused::times_walltime>(instance) = extractedData.times.walltime;
      boost::fusion::at_key<MPQCDataFused::times_cputime>(instance) = extractedData.times.cputime;
      boost::fusion::at_key<MPQCDataFused::times_flops>(instance) = extractedData.times.flops;
      MPQCData_Time_fused.push_back(instance);
    }

    // create a vector of all job ids
    std::vector<JobId_t> jobids(results.size(), JobId::IllegalJob);
    std::transform(results.begin(), results.end(), jobids.begin(),
        boost::bind(&FragmentResult::getId,
            boost::bind(&FragmentResult::ptr::operator->, _1)));

    // sum up and print energies
    boost::mpl::for_each<MPQCDataEnergyVector_t>(
        printSum<MPQCDataEnergyMap_t>(
            subsetmap,
            MPQCData_Energy_fused,
            jobids,
            container->getContainer(),
            MatrixNrLookup)
        );
    // TODO: Histogram still needs cstor taking just vector<double>


    // TODO: We need to put forces into specific class and implement operators
//    // sum up and print forces
//    boost::mpl::for_each<MPQCDataForceVector_t>(
//        printSum<MPQCDataForceMap_t>(
//            subsetmap,
//            MPQCData_Force_fused,
//            jobids,
//            container->getContainer(),
//            MatrixNrLookup)
//        );

    // TODO: Times should just be summed not orthogonalized
//    // sum up and print times
//    boost::mpl::for_each<MPQCDataTimeVector_t>(
//        printSum<MPQCDataTimeMap_t>(
//            subsetmap,
//            MPQCData_Time_fused,
//            jobids,
//            container->getContainer(),
//            MatrixNrLookup)
//        );
  }

  // combine all found data
  if (!KeySet.ParseManyBodyTerms()) return false;

  if (!EnergyFragments.AllocateMatrix(Energy.Header, Energy.MatrixCounter, Energy.RowCounter, Energy.ColumnCounter)) return false;
  if (!ForceFragments.AllocateMatrix(Force.Header, Force.MatrixCounter, Force.RowCounter, Force.ColumnCounter)) return false;

  if(!Energy.SetLastMatrix(0., 0)) return false;
  if(!Force.SetLastMatrix(0., 2)) return false;

  for (int BondOrder=0;BondOrder<KeySet.Order;BondOrder++) {
    // --------- sum up energy --------------------
    LOG(1, "INFO: Summing energy of order " << BondOrder+1 << " ...");
    if (!EnergyFragments.SumSubManyBodyTerms(Energy, KeySet, BondOrder)) return false;
    if (!Energy.SumSubEnergy(EnergyFragments, NULL, KeySet, BondOrder, 1.)) return false;

    // --------- sum up Forces --------------------
    LOG(1, "INFO: Summing forces of order " << BondOrder+1 << " ...");
    if (!ForceFragments.SumSubManyBodyTerms(Force, KeySet, BondOrder)) return false;
    if (!Force.SumSubForces(ForceFragments, KeySet, BondOrder, 1.)) return false;
  }

  // for debugging print resulting energy and forces
  LOG(1, "INFO: Resulting energy is " << Energy.Matrix[ FragmentCounter ][0][0]);
  std::stringstream output;
  for (int i=0; i< Force.RowCounter[FragmentCounter]; ++i) {
    for (int j=0; j< Force.ColumnCounter[FragmentCounter]; ++j)
      output << Force.Matrix[ FragmentCounter ][i][j] << " ";
    output << "\n";
  }
  LOG(1, "INFO: Resulting forces are " << std::endl << output.str());

  return true;
}


void RunService(
    boost::asio::io_service &io_service,
    std::string message)
{
  message = std::string("io_service: ") + message;
  io_service.reset();
  Info info(message.c_str());
  io_service.run();
}

void requestIds(
    FragmentController &controller,
    const FragmentationFragmentationAutomationAction::FragmentationFragmentationAutomationParameters &params,
    const size_t numberjobs)
{
  controller.requestIds(params.host.get(), params.port.get(), numberjobs);
}

bool createJobsFromFiles(
    FragmentController &controller,
    const FragmentationFragmentationAutomationAction::FragmentationFragmentationAutomationParameters &params,
    const std::vector< boost::filesystem::path > &jobfiles)
{
  std::vector<FragmentJob::ptr> jobs;
  for (std::vector< boost::filesystem::path >::const_iterator iter = jobfiles.begin();
      iter != jobfiles .end(); ++iter) {
    const std::string &filename = (*iter).string();
    if (boost::filesystem::exists(filename)) {
      const JobId_t next_id = controller.getAvailableId();
      LOG(1, "INFO: Creating MPQCCommandJob with filename'"
          +filename+"', and id "+toString(next_id)+".");
      parsejob(jobs, params.executable.get().string(), filename, next_id);
    } else {
      ELOG(1, "Fragment job "+filename+" does not exist.");
      return false;
    }
  }
  controller.addJobs(jobs);
  controller.sendJobs(params.host.get(), params.port.get());
  return true;
}

void WaitforResults(
    boost::asio::io_service &io_service,
    FragmentController &controller,
    const FragmentationFragmentationAutomationAction::FragmentationFragmentationAutomationParameters &params,
    const size_t NoExpectedResults
    )
{
  size_t NoCalculatedResults = 0;
  while (NoCalculatedResults != NoExpectedResults) {
    // wait a bit
    boost::asio::deadline_timer timer(io_service);
    timer.expires_from_now(boost::posix_time::milliseconds(500));
    timer.wait();
    // then request status
    controller.checkResults(params.host.get(), params.port.get());
    RunService(io_service, "Checking on results");

    const std::pair<size_t, size_t> JobStatus = controller.getJobStatus();
    LOG(1, "INFO: #" << JobStatus.first << " are waiting in the queue and #" << JobStatus.second << " jobs are calculated so far.");
    NoCalculatedResults = JobStatus.second;
  }
}


Action::state_ptr FragmentationFragmentationAutomationAction::performCall() {
  boost::asio::io_service io_service;
  FragmentController controller(io_service);

  // TODO: Have io_service run in second thread and merge with current again eventually

  // Phase One: obtain ids
  std::vector< boost::filesystem::path > jobfiles = params.jobfiles.get();
  requestIds(controller, params, jobfiles.size());
  RunService(io_service, "Requesting ids");

  // Phase Two: create and add jobs
  if (!createJobsFromFiles(controller, params, jobfiles))
    return Action::failure;
  RunService(io_service, "Adding jobs");

  // Phase Three: calculate result
  WaitforResults(io_service, controller, params, jobfiles.size());

  // Phase Three: get result
  controller.receiveResults(params.host.get(), params.port.get());
  RunService(io_service, "Phase Four");

  // Final phase: print result
  {
    LOG(1, "INFO: Parsing fragment files from " << params.path.get() << ".");
    std::vector<FragmentResult::ptr> results = controller.getReceivedResults();
    printReceivedMPQCResults(
        results,
        params.path.get(),
        getNoAtomsFromAdjacencyFile(params.path.get()));
  }
  size_t Exitflag = controller.getExitflag();

  return (Exitflag == 0) ? Action::success : Action::failure;
}

Action::state_ptr FragmentationFragmentationAutomationAction::performUndo(Action::state_ptr _state) {
  return Action::success;
}

Action::state_ptr FragmentationFragmentationAutomationAction::performRedo(Action::state_ptr _state){
  return Action::success;
}

bool FragmentationFragmentationAutomationAction::canUndo() {
  return false;
}

bool FragmentationFragmentationAutomationAction::shouldUndo() {
  return false;
}
/** =========== end of function ====================== */